Yes and no. I've seen the same sht happen twice, in the same year.

First in the telco crash of early 2001, when telco shares collapsed, in no small part dues to the backbone fiber rollouts done for the nascent internet (as well as many telco customers became illiquid bidding for 3G spectrum), said backbone sat unussed for years (and the spectrum too), but, at least, the fiber did not depreciate at an alarming rate...

But then, by the middle of that same year 2001, the Internet bubble bust. Untold ammounts of datacenter capacity went unused, cooling and motorgenerators dimensioned for big loads faced significantly smaller loads, racks upon racks of rapidly depreciating servers went unused, and untold ammounts of warehouses and perishable inventory (think pets.com 's pet food) sat there, again, unused.

Well, the world recovered, the internet provet to be worthy, the telco fiber and spectrum saw a lot of use latter on...

The same will happen with AI... just pray the AI bubble deflates, and does not POP! (Pop! goes the world).

Linus Torvalds, Greg H-K and the Mozilla team are singign the praises on AI for software maintenance. And now a 19 year old FOSS grapghics driver is still getting software improvements thanks to AI!

And yet some zealots are saying that AI has no use whatsoever...

You know what? More than one thing can be true at once.

Yes, is true that AI is not a panacea that will replace every single coder/white collar job.

Yes, is true that judiciously used, AI can be extremely useful for many task inside many a job description, including sw development.

The world is not black and white, or even shades of gray, at least for an electronics engineer like me is not only in technocolor, but in even more wavelenghts, and polarized horizontally, vertically and elliptically to boot :-P

In developed countries, where people and companies can buy the hardware (say, a cluster of ryzen strix halos) it makes no sense to pay monthly for meta's AI. But in countries where people and companies can not bear the upfront costs? This actually makes sense.

Read again the list of pilot countries in light of the previous paaragraph.

Full disclosure: i live in LatAm, but not in Bolivia

Not only they crave it, but also, the job itself demands it.
And part of the attention is the severity of the Bug, with security bugs with working exploit code being the "best-est".

So, In the same address, Torvads asked security researchers to not publish exploit code, but this goes against the incentive structure of security research including payment.

Luckily, fixing that problem is easy. Linux is taling about a (current) private security mailing list, and a (future) Public list.

Well, if you are a security researcher, subbmit your bug to both lists, first to the private mailing list, with the example exploit code, then to the public list, sans the exploit code, but with an adendum that says "exploit code avaialable in the private security list under security bug report # xx.yy.zz". When the security hole you reported is patched, and the details of the private mailing list become public, and the exploit code is shown to work, the infosec researcher (if s/he responded and did follow-up work) will be dully cretited, which is nice and works for everyone.

JM2C YMMV

The synology i bought in 2016 had no hardware media codecs or GPU. So the plex pass had no inherent advabtage.

Also, it came with VideoStation. At the time plex did some things that videostation did not, and videostation did things that plex did not (in particular, folder view, and android app at no extra cost), so they complemented each other nicely. I used both.

Now, if you re-introduce video station, it is half-broken, and plex has deviated from local media streaming to streaming a la carte + live tv + games + rentals...

I'll keep using Plex becuase i am familiar with it (warts and all) until 2028 when my NAS is due for a change... Then, jellyfin it is.

To save vertical space. But bigger and more pixel rich monitors took care of that.

My funniest use of a vertical taskbar was for a cousin who broke the upper left corner of his laptop's LCD. Vertical taskbar + Wider taskbar saved that laptop from the scrap heap, as the crack only obscured the start button and some pinned crap.

I did not caught on to Sysadmins moving the taskbars around so that when they remoted into machines, they knew, at a glance, exactly where were they (taskbar on the left: My machine, taskbar on the bottom, user machine, taskbar on the right: server, taskbar on the top, VERY SPECIAL MACHINE HANDLE WITH CARE).

But these are niche uses of a movable taskbar. For office drones (which is the bread and butter of Windows client editions), the best way to go is to have the taskbar in a fixed location, ease of training, ease of documentation, ease of support, ease for two people huddled in front of one screen.

I understand why microsoft did it. And I understand why they are backpedaling.

The Williams Company strung fiber optic cables inside decomisiones Gas Pipes, that was Wiltel. first iteration bought by LDDS, second one bought by Level 3
https://en.wikipedia.org/wiki/...

Quest laid down fiber alongside train right of way, using a special plough moved by a locomotive
https://en.wikipedia.org/wiki/...

Ditto for laying bahaul fibre in the eguts/sewers of large cities. Or using the actual sewer pipes to bring the access fiber to buildings or houses...

If the rights of ways are aquired for something else already, laying the fiber is easy and cheap, and a nice way to earn additional revenues on your existing rights of way

That seminal paper was published in 2004. Yes here we are, 22 years later
Some people were asleep at the wheel.

AI training need HUUUUGE datacenters. But AI inference may benefit from these smaller datacenters.

Ditto if you can move classic cloud workloads and associated gear (SaaS, PaaS, IaaS) to these smaller datacenters, freeing up your biggest datacenters for AI training.

I proposed to do something similar to that a while ago, using container data centers. But this may work even better.

From TFS:

I guess this guy is either hoping for, or actually has other investors ready to provide the other U$D 16 milliard. Probably, those investors will come forward once the deal clears some other hurdles.

When I studied French in Laussane in 1997, you were fresh out of an EU joining referendum. You said no. At the time I tought that was a bad decistion. While I still think that being a part of a greater whole is better than beain a small independent part, it is now evident that the EU as it currently is, does not gel well (or is downright incompatible) with what Swiss is and always has been...

JM2C
YMMV

They are efficient. Never felt the need to privatize...

What do the swiss have that the belgians lack?

"with the current generation of young programmers. They clearly do not know the difference between an operating system and applications. Nobody should be trying to add AI to Windows, or to Linux, or to any other OS. The OS is supposed to add a layer of abstraction to the platform, so applications can be written and then run on multiple systems with hardware differences. The OS is supposed to allocate resources to applications. The modern OS is supposed to allow multiple applications to run at the same time or appear to run at the same time using some combination of cores and time-slicing. If any operating system is having problems doing these things (the basics) then programmers should be improving whichever element is not up to par."

I call bull. Older versions of unix came with apps. An editor (vi), network utilities (mail, gopher, ping, traceroute), utilities like grep, interpreted languages like awk...

Is logical that, as the OS went graphical, vi gave way to notepad++ (recently released for linux).

Remeber all the controversy in the mid '90s about if the browser should ship with the OS or not. Nowadays you would not accept a desktop OS without a browser out of the box.

AI is the next iteration of this. The desktop OS HAS to have AI out of the box. If you want to disable or replace it, so be it, but an enterprise-y desktop OS like ubuntu HAS to have AI out of the box, for the convenience of the corpos that buy support and therefore pay the development that benefits us free users

A cursory read of the comments show that the people on these parts want no AI in their Ubuntu. Which is not what you said

And, on a personal note, i do not think is wise for a generalist and corporate distro like ububtu to ship sans AI integrations in the year of our lord 2026. As long as canonical has a light touch and do not ram it though our throats like windows, everythig should be ok

I know that, you know that, we both know that. But a cursory read of the other comments shows that most people in these parts did not get the memo, an pretend that every single people using linux get a no-ai experience, and has the l337 skillzzz to install huggingface and do the model integration on all the relevant parts of the OS themselves....

I fail to see how accountants, administrators, video editors, photographers and biologists, just to name a few, will acomplish that.

But then again, what does an OpenStack trainer like me can possibly know about the l337 skillzzz of normies to install and deeply integrate AI on a non-AI OS.....